How do you partition the dataset into train/val/test?

[ArchNew]

I downloaded the dataset. But it seems the json file of your dataset doesn't indicate the partition of the dataset. I have no idea which part is the training set and which parts are the val and test sets?